1 Loading data and packages

library(readxl)
library(readr)
library(CTexploreR)
library(Vennerable)
library(biomaRt)
library(tidyverse)
library(SummarizedExperiment)
library(UpSetR)

Gene names/synonyms required for databases cleaning

ensembl <- biomaRt::useDataset("hsapiens_gene_ensembl", useMart("ensembl"))
attributes_vector <- c("ensembl_gene_id", "external_gene_name",
                       "external_synonym", "gene_biotype",
                       "chromosome_name", "band", "start_position", "end_position",
                       "strand")
ensembl_gene_synonym <- as_tibble(getBM(attributes = attributes_vector, mart = ensembl))

ensembl_gene_synonym <- ensembl_gene_synonym %>%
  mutate(external_synonym = sub(pattern = "ORF", external_synonym, 
                                replacement = "orf"))

attributes_vector <- c("ensembl_gene_id", "external_gene_name")
ensembl_gene_names <- as_tibble(getBM(attributes = attributes_vector, mart = ensembl))

attributes_vector <- c("external_gene_name",
                       "external_synonym")
gene_synonym <- as_tibble(getBM(attributes = attributes_vector, mart = ensembl))
load("../CTdata/eh_data/all_genes.rda")
load("../CTdata/eh_data/CT_genes.rda")

CT_and_CTP_genes <- CT_genes
CT_genes <- (filter(CT_genes, CT_gene_type == "CT_gene"))


upset_text_size <- c(1.5, 1.5, 1.5, 1.5, 1.8, 2)
 #c(intersection size title, intersection size tick labels, set size title, 
 # set size tick labels, set names, numbers above bars)

2 Database comparison

2.1 Lists cleaning

CT lists from other databases have been checked (using GTEx and our GTEx_expression() funtion and GeneCards) in order to remove duplicated gene names or deprecated ones and allow comparison between databases.

2.1.1 CTdatabase

Online list copied in a csv file, several lists exist so we combined them.

We checked gene names that were a concatenation of two genes (choice using biomaRt synonyms to get the official one), checked which ones had the right names, removed duplicated genes, verified lost genes and added back those that should be there.

CTdatabase <- read_delim("data/CTdatabase1.csv", delim = ";", 
                         escape_double = FALSE, trim_ws = TRUE)
colnames(CTdatabase) <- c("Family", "Gene_Name", "Chromosomal_localization",
                          "CT_identifier")
CTdatabase_bis <- read_csv2("data/CTdatabase2.csv")
CTdatabase <- left_join(CTdatabase, CTdatabase_bis, 
                        by = c("Gene_Name" = "Gene_Symbol"))


CTdatabase_single <- CTdatabase %>%
  mutate(Gene_Name = sub(pattern = "/.*$", Gene_Name, replacement = ""))
CTdatabase_single <- CTdatabase_single %>%
  mutate(Gene_Name = sub(pattern = ",.*$", Gene_Name, replacement = ""))


CTdatabase_official_names <- 
  unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, 
                       external_gene_name)) %>%
  filter(external_gene_name %in% CTdatabase_single$Gene_Name) %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA)
CTdatabase_synonym <- 
  ensembl_gene_synonym %>%
  filter(external_synonym %in% CTdatabase_single$Gene_Name) %>%
  mutate(Gene_Name = external_synonym) %>%
  dplyr::select(ensembl_gene_id, external_gene_name, Gene_Name, external_synonym)
CTdatabase_cleaned <- 
  rbind(CTdatabase_official_names, CTdatabase_synonym) %>% 
  left_join(CTdatabase_single)


duplicated_genes <- CTdatabase_cleaned$Gene_Name[duplicated(CTdatabase_cleaned$Gene_Name)]
bad_ids <- ensembl_gene_synonym %>%
  filter(external_gene_name %in% duplicated_genes | external_synonym %in% duplicated_genes) %>%
  filter(chromosome_name %in% grep(pattern = "H", x = chromosome_name, value = TRUE)) %>%
  pull(ensembl_gene_id)
CTdatabase_cleaned <- CTdatabase_cleaned %>%
  dplyr::filter(!ensembl_gene_id %in% bad_ids)
CTdatabase_cleaned <- CTdatabase_cleaned %>%
  filter(!ensembl_gene_id == "ENSG00000052126")
CTdatabase_cleaned <- CTdatabase_cleaned %>% 
  filter(!(ensembl_gene_id == "ENSG00000183305" & Gene_Name == "MAGEA2"))
CTdatabase_cleaned <- CTdatabase_cleaned %>% 
  filter(!ensembl_gene_id == "ENSG00000204648")
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "CSAG3B")
CTdatabase_cleaned[CTdatabase_cleaned$Gene_Name == "CSAG2", "external_synonym"] <- "CSAG3B"
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "CT45A4")
CTdatabase_cleaned[CTdatabase_cleaned$Gene_Name == "CT45A3", "external_synonym"] <- "CT45A4"
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "LAGE-1b")
CTdatabase_cleaned[CTdatabase_cleaned$Gene_Name == "CTAG2", "external_synonym"] <- "LAGE-1b"
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "CT16.2")
CTdatabase_cleaned[CTdatabase_cleaned$Gene_Name == "PAGE5", "external_synonym"] <- "CT16.2"
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "SPANXB2")
CTdatabase_cleaned[CTdatabase_cleaned$Gene_Name == "SPANXB1", "external_synonym"] <- "SPANXB2"
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "SPANXE")
CTdatabase_cleaned[CTdatabase_cleaned$Gene_Name == "SPANXD", "external_synonym"] <- "SPANXE"
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "XAGE1C")
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "XAGE1D")
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "XAGE1E")
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "XAGE2B")
CTdatabase_cleaned[CTdatabase_cleaned$Gene_Name == "XAGE2", "external_synonym"] <- "XAGE2B"
CTdatabase_cleaned <- filter(CTdatabase_cleaned, Gene_Name != "CTAGE-2")
CTdatabase_cleaned[CTdatabase_cleaned$Gene_Name == "CTAGE1", "external_synonym"] <- "CTAGE-2"


CTdatabase_cleaned <- ensembl_gene_synonym %>%
  mutate(Gene_Name = external_synonym) %>%
  filter(external_synonym == "CXorf61") %>%
  dplyr::select(ensembl_gene_id, external_gene_name, Gene_Name, external_synonym) %>%
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "Cxorf61", 
                    c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)
CTdatabase_cleaned <- unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, external_gene_name)) %>%
  filter(external_gene_name == "CCNA1") %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA) %>% 
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "cyclin A1", 
                          c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)
CTdatabase_cleaned <- unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, external_gene_name))%>%
  filter(external_gene_name == "GOLGA6L2") %>%
  filter(ensembl_gene_id == "ENSG00000174450") %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA) %>% 
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "GOLGAGL2 FA", 
                          c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)
CTdatabase_cleaned <- unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, external_gene_name))%>%
  filter(external_gene_name == "LYPD6B") %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA) %>% 
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "LOC130576", 
                          c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)
CTdatabase_cleaned <- unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, external_gene_name))%>%
  filter(external_gene_name == "CT62") %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA) %>% 
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "LOC196993", 
                          c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)
CTdatabase_cleaned <- unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, external_gene_name))%>%
  filter(external_gene_name == "CT75") %>%
  filter(ensembl_gene_id == "ENSG00000291155") %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA) %>% 
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "LOC440934", 
                          c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)
CTdatabase_cleaned <- unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, external_gene_name))%>%
  filter(external_gene_name == "LINC01192") %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA) %>% 
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "LOC647107", 
                          c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)
CTdatabase_cleaned <- unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, external_gene_name))%>%
  filter(external_gene_name == "TSPY1") %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA) %>% 
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "LOC728137", 
                          c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)
CTdatabase_cleaned <- unique(dplyr::select(ensembl_gene_synonym, ensembl_gene_id, external_gene_name))%>%
  filter(external_gene_name == "SSX2B") %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA) %>% 
  cbind(CTdatabase_single[CTdatabase_single$Gene_Name == "SSX2b", 
                          c("Family", "Chromosomal_localization", "CT_identifier", "Classification")]) %>% 
  rbind(CTdatabase_cleaned)

2.1.2 Jamin’s list

Excel file coming from supplemental data.

Jamin_core_CT <- read_excel("data/Jamin_core_CT.xlsx")
Jamin_core_CT[Jamin_core_CT$Gene == "KIAA1211", "Gene"] <- "CRACD"
Jamin_core_CT[Jamin_core_CT$Gene == "CXorf67", "Gene"] <- "EZHIP"

2.1.3 Wang’s CTatlas

Excel file coming from supplemental data.

Wang_CT <- read_excel("data/Wang_Suppl_Data_3.xlsx", 
    sheet = "Supplementary Data 3B", skip = 1)
colnames(Wang_CT)[1] <- "ensembl_gene_id"

Wang_CT <- ensembl_gene_names %>% 
  filter(ensembl_gene_id %in% Wang_CT$ensembl_gene_id) %>%
  right_join(Wang_CT)

Wang_CT[Wang_CT$ensembl_gene_id == "ENSG00000181013", "external_gene_name"] <- "C17orf47"
Wang_CT[Wang_CT$ensembl_gene_id == "ENSG00000204293", "external_gene_name"] <- "OR8B2"
Wang_CT[Wang_CT$external_gene_name == "", "external_gene_name"] <- "RNASE11"
Wang_CT[Wang_CT$external_gene_name == "CHCT1", "external_gene_name"] <- "C17orf64"
Wang_CT[Wang_CT$external_gene_name == "PRSS40A", "external_gene_name"] <- "TISP43"
Wang_CT[Wang_CT$external_gene_name == "TEX56P", "external_gene_name"] <- "C6orf201"
Wang_CT[Wang_CT$external_gene_name == "SLC25A51P4", "external_gene_name"] <- "RP11-113D6.10"
Wang_CT[Wang_CT$external_gene_name == "TCP10L3", "external_gene_name"] <- "TCP10"
Wang_CT[Wang_CT$external_gene_name == "SCAND3", "external_gene_name"] <- "ZBED9"

2.1.4 Carter’s list

Carter_CT_list <- read_table("data/Carter_CT_list.txt", skip = 1)
Carter_CT <- Carter_CT_list[Carter_CT_list$CT_Expression,]

Carter_CT[Carter_CT$Gene == "ENSG00000261649", "Gene_Name"] <- "GOLGA6L7"
Carter_CT[Carter_CT$Gene == "ENSG00000239620", "Gene_Name"] <- "PRR20G"
Carter_CT[Carter_CT$Gene == "ENSG00000168148", "Gene_Name"] <- "H3-4"
Carter_CT[Carter_CT$Gene == "ENSG00000204296", "Gene_Name"] <- "TSBP1"
Carter_CT[Carter_CT$Gene == "ENSG00000180219", "Gene_Name"] <- "GARIN6"
Carter_CT[Carter_CT$Gene == "ENSG00000172717", "Gene_Name"] <- "GARIN2"
Carter_CT[Carter_CT$Gene == "ENSG00000174015", "Gene_Name"] <- "CBY2"
Carter_CT[Carter_CT$Gene == "ENSG00000224960", "Gene_Name"] <- "PPP4R3C"

2.1.5 Burggeman’s list

Excel file from supplemental data.

Bruggeman_data <- read_excel("data/Bruggeman_suppl_data.xlsx", skip = 1,
                           sheet = "1D")

Bruggeman_official_names <- gene_synonym %>% 
  dplyr::select(external_gene_name) %>% 
  unique() %>% 
  filter(external_gene_name %in% Bruggeman_data$Gene) %>%
  mutate(Gene_Name = external_gene_name) %>%
  mutate(external_synonym = NA)

Bruggeman_synonym <- gene_synonym %>%
  filter(external_synonym %in% Bruggeman_data$Gene) %>%
  mutate(Gene_Name = external_synonym) %>%
  dplyr::select(external_gene_name, Gene_Name, external_synonym)

Bruggeman_synonym <- Bruggeman_synonym[-which(Bruggeman_synonym$Gene_Name %in% 
                          Bruggeman_official_names$Gene_Name),]

Bruggeman_CT <- rbind(Bruggeman_official_names, Bruggeman_synonym)

lost <- Bruggeman_data[which(!Bruggeman_data$Gene %in% c(Bruggeman_CT$Gene_Name)), "Gene"]
colnames(lost) <- "external_gene_name"
lost$Gene_Name <- rep(NA, nrow(lost))
lost$external_synonym <- rep(NA, nrow(lost))

lost[lost$external_gene_name == "C21orf59", "Gene_Name"] <- "CFAP298"
lost[lost$external_gene_name == "C11orf57", "Gene_Name"] <- "NKAPD1"
lost[lost$external_gene_name == "C7orf55", "Gene_Name"] <- "FMC1"
lost[lost$external_gene_name == "C10orf12", "Gene_Name"] <- "LCOR"
lost[lost$external_gene_name == "RPL19P12", "Gene_Name"] <- "RPL19P12"
lost[lost$external_gene_name == "C16orf59", "Gene_Name"] <- "TEDC2"
lost[lost$external_gene_name == "TTTY15", "Gene_Name"] <- "USP9Y"
lost[lost$external_gene_name == "C17orf53", "Gene_Name"] <- "HROB"
lost[lost$external_gene_name == "C1orf112", "Gene_Name"] <- "FIRRM"
lost[lost$external_gene_name == "C12orf66", "Gene_Name"] <- "KICS2"
lost[lost$external_gene_name == "C9orf84", "Gene_Name"] <- "SHOC1"
lost[lost$external_gene_name == "C10orf25", "Gene_Name"] <- "ZNF22-AS1"
lost[lost$external_gene_name == "C20orf197", "Gene_Name"] <- "LINC02910"
lost[lost$external_gene_name == "C3orf67", "Gene_Name"] <- "CFAP20DC"
lost[lost$external_gene_name == "C8orf37", "Gene_Name"] <- "CFAP418"
lost[lost$external_gene_name == "C22orf34", "Gene_Name"] <- "MIR3667HG"
  
Bruggeman_CT <- rbind(Bruggeman_CT, lost) 

missing_Bruggeman <- c("BMS1P4", "ADAM6", "ANXA2P3", "ARHGAP11B", "DPY19L2P2", 
                       "HLA-L", "PA2G4P4", "PIPSL", "PRKY", "YBX3P1", 
                       "RPL23AP53", "UQCRBP1", "RPL23P8", "MRS2P2", "PIN4P1", 
                       "SLC6A10P", "GUSBP2", "PPIEL", "LRRC37BP1", "MSL3P1", 
                       "PLEKHA8P1", "STAG3L1", "TCAM1P", "ZNF702P", "ZNF815P", 
                       "ATP6AP1L", "RPL21P44", "SEC14L1P1", "ZNF876P", 
                       "RPLP0P2", "FAM86JP", "FAM175A", "LACE1", "ATP5EP2", 
                       "WDR92", "TCTE3", "METTL20", "KIAA2022", "ZNRD1-AS1", 
                       "SGOL1", "FAM35DP", "MTL5", "TMEM14E", "MLLT4-AS1", 
                       "CCDC173", "KIAA1524", "WDR78", "LINC00476", "LYRM5", 
                       "HILS1", "CASC5", "KIAA1919", "CTAGE5", "FAM188B", 
                       "TMEM194B", "FAM122C", "PPP1R2P3", "KIAA0391", "SGOL2", 
                       "FAM19A3", "ZNF788", "RPL19P12", "FIRRM")

external_names_to_keep <- gene_synonym %>%
   filter(external_synonym %in% missing_Bruggeman) %>%
   filter(!external_gene_name %in% c("ATP5F1EP2", "POLR1HASP", "SHLD2P3", 
                                   "TMEM14EP", "H1-9P", "ZNF788P")) %>% 
   mutate(Gene_Name = external_gene_name)
 
Bruggeman_CT[Bruggeman_CT$external_synonym %in% 
                    external_names_to_keep$external_synonym, 
                  "Gene_Name"] <- external_names_to_keep$Gene_Name

Bruggeman_CT <- Bruggeman_CT %>% 
  dplyr::select(Gene_Name)

2.2 CTexploreR data for selection pipeline

To characterise the differences between our database and other, we need the category we created in CTexploreR. For this, we have the object all_genes in CTdata that contains the CT analysis for all genes. More info in

Hereunder is what we used for our selection pipeline (coming from make_all_genes_prelim.R and 130_make_all_genes_and_CT_genes.R in CTdata).

all_genes

From there, we filtered based on the testis_specificity (“testis_specific”, which is based on expression in health tissue and scRNA seq info from HPA), CCLE_category (“activated”) and TCGA_category (“activated” or “multimapping_issue”) to have our CT genes. Then, when wanting to validate TSS manually, we realised that for some genes, reads were not properly aligned to exons which might reflect a poorly defined transcription in these regions and are hence likely unreliable.

Some genes were also characterized as Cancer-Testis preferential genes when testis specificity was less stringent

2.3 CTexploreR VS CTdatabase

CTdatabase_ours <- Venn(list(CTdatabase = CTdatabase_cleaned$external_gene_name,
                             CTexploreR = CT_genes$external_gene_name))
gp <- VennThemes(compute.Venn(CTdatabase_ours))
gp[["Face"]][["11"]]$fill <-  "mistyrose"
gp[["Face"]][["01"]]$fill <-  "darkseagreen1"
gp[["Face"]][["10"]]$fill <-  "lightsteelblue1"
gp[["Set"]][["Set1"]]$col <-  "paleturquoise4"
gp[["Set"]][["Set2"]]$col <-  "darkseagreen4"
gp[["SetText"]][["Set1"]]$col <-  "paleturquoise4"
gp[["SetText"]][["Set2"]]$col <-  "darkseagreen4"
plot(CTdatabase_ours, gp = gp)

We find 29.0322581 % of CTdatabase in CTexploreR, which is 40.9090909% of our database.

Lost genes analysis

CTdatabase_lost <- all_genes %>%
  filter(external_gene_name %in% CTdatabase_ours@IntersectionSets[["10"]])

# 9 Genes are lost because not in any database

table(CTdatabase_lost$testis_specificity)
## 
## not_testis_specific testis_preferential     testis_specific 
##                  70                  54                  41
table(CTdatabase_lost$CT_gene_type)
## 
## CTP_gene    other 
##       15      150
table(CTdatabase_lost$not_detected_in_somatic_HPA)
## 
## FALSE  TRUE 
##    51   106
table(CTdatabase_lost$TCGA_category)
## 
##          activated              leaky multimapping_issue      not_activated 
##                 34                 83                 25                 23
table(CTdatabase_lost$CCLE_category)
## 
##     activated         leaky not_activated 
##            81            33            51
table(CTdatabase_lost$TCGA_category, CTdatabase_lost$CCLE_category)
##                     
##                      activated leaky not_activated
##   activated                 20     0            14
##   leaky                     41    33             9
##   multimapping_issue        15     0            10
##   not_activated              5     0            18
CTdatabase_lost_upset <- 
  list(`Not testis specific` = 
         filter(CTdatabase_lost,
                testis_specificity != "testis_specific")$external_gene_name,
       `Not tumour activated` = 
         filter(CTdatabase_lost,
                (TCGA_category != "activated" &
                  TCGA_category != "multimapping_issue")|
                   CCLE_category != "activated")$external_gene_name,
       `CT preferential` =
         filter(CTdatabase_lost,
                CT_gene_type == "CTP_gene")$external_gene_name)

upset_CTdatabase <- fromList(CTdatabase_lost_upset)

upset(upset_CTdatabase,
      text.scale = upset_text_size)

75.1515152 % of these genes are not testis specific.

However 15 of these lost genes are flagged as Cancer-Testis preferential in our analysis.

87.8787879 % are not properly activated in tumors and/or cancer cell lines.

In their analysis, they had characterised gene specificity, some being not available, not found in testis, testis-restricted, testis-selective and testis/brain-restricted. Let’s see how the lost genes qualify as they didn’t mention those were strictly testis specific.

CTdatabase_cleaned %>% 
  filter(external_gene_name %in% CTdatabase_lost$external_gene_name) %>% 
  pull(Classification) %>% 
  table()
## .
##           not available     not found in testis       testis-restricted 
##                      90                       2                      15 
##        testis-selective testis/brain-restricted 
##                      51                       7

We can see that most of them had no info or were testis-selective (I couldn’t find on website or paper how they selected categories).

2.4 CTexploreR VS omics databases

core_ours <- Venn(list(Jamin = Jamin_core_CT$Gene, 
                       CTexploreR = CT_genes$external_gene_name))

Wang_ours <- Venn(list(Wang = Wang_CT$external_gene_name, 
                       CTexploreR = CT_genes$external_gene_name))

Carter_ours <- Venn(list(Carter_CT = Carter_CT$Gene_Name, 
                         CTexploreR = CT_genes$external_gene_name))

Bruggeman_ours <- Venn(list(Bruggeman = Bruggeman_CT$Gene_Name,
                            CTexploreR = CT_genes$external_gene_name))

gene_list <- list(CTexploreR = CT_genes$external_gene_name,
                  Carter = Carter_CT$Gene_Name,
                  Jamin = Jamin_core_CT$Gene, 
                  CTatlas = Wang_CT$external_gene_name,
                  Bruggeman = Bruggeman_CT$Gene_Name)

upset_omics <- fromList(gene_list)
upset(upset_omics)

4 in all, 60 in at least 3 databases

Lost genes analysis

plot(core_ours, gp = gp)

Jamin_lost <- all_genes %>%
  filter(external_gene_name %in% core_ours@IntersectionSets[["10"]])

table(Jamin_lost$testis_specificity)
## 
## not_testis_specific testis_preferential     testis_specific 
##                  70                  20                   2
table(Jamin_lost$CT_gene_type)
## 
## CTP_gene    other 
##       10       82
table(Jamin_lost$not_detected_in_somatic_HPA)
## 
## FALSE  TRUE 
##    44    48
table(Jamin_lost$TCGA_category)
## 
##          activated              leaky multimapping_issue      not_activated 
##                  8                 63                 17                  4
table(Jamin_lost$CCLE_category)
## 
##     activated         leaky not_activated 
##            49            39             4
table(Jamin_lost$TCGA_category, Jamin_lost$CCLE_category)
##                     
##                      activated leaky not_activated
##   activated                  8     0             0
##   leaky                     22    39             2
##   multimapping_issue        17     0             0
##   not_activated              2     0             2
Jamin_lost_upset <- 
  list(`Not testis specific` = 
         filter(Jamin_lost,
                testis_specificity != "testis_specific")$external_gene_name,
       `Not tumour activated` = 
         filter(Jamin_lost,
                (TCGA_category != "activated" &
                  TCGA_category != "multimapping_issue")|
                   CCLE_category != "activated")$external_gene_name,
       `CT preferential` =
         filter(Jamin_lost,
                CT_gene_type == "CTP_gene")$external_gene_name)

upset_Jamin <- fromList(Jamin_lost_upset)
upset(upset_Jamin,
      text.scale = upset_text_size)

We find 23.2 % of CTdatabase in CTexploreR, which is 16.4772727 % of our database.

97.826087% of these genes are not testis specific.

However 10 of these lost genes are flagged as Cancer-Testis preferential in our analysis.

91.3043478 % are not properly activated in tumors and/or cancer cell lines.

plot(Wang_ours, gp = gp)

Wang_lost <- all_genes %>%
  filter(external_gene_name %in% Wang_ours@IntersectionSets[["10"]])


table(Wang_lost$testis_specificity)
## 
## not_testis_specific testis_preferential     testis_specific 
##                 335                 347                 227
table(Wang_lost$CT_gene_type)
## 
## CTP_gene    other 
##       40      869
table(Wang_lost$not_detected_in_somatic_HPA)
## 
## FALSE  TRUE 
##   254   625
table(Wang_lost$TCGA_category)
## 
##          activated              leaky multimapping_issue      not_activated 
##                188                473                 66                182
table(Wang_lost$CCLE_category)
## 
##     activated         leaky not_activated 
##           342           205           362
table(Wang_lost$TCGA_category, Wang_lost$CCLE_category)
##                     
##                      activated leaky not_activated
##   activated                 88     3            97
##   leaky                    188   200            85
##   multimapping_issue        34     0            32
##   not_activated             32     2           148
Wang_lost_upset <- 
  list(`Not testis specific` = 
         filter(Wang_lost,
                testis_specificity != "testis_specific")$external_gene_name,
       `Not tumour activated` = 
         filter(Wang_lost,
                (TCGA_category != "activated" &
                  TCGA_category != "multimapping_issue")|
                   CCLE_category != "activated")$external_gene_name,
       `CT preferential` =
         filter(Wang_lost,
                CT_gene_type == "CTP_gene")$external_gene_name)

upset_Wang <- fromList(Wang_lost_upset)
upset(upset_Wang,
      text.scale = upset_text_size)

We find 9.0284593 % of CTdatabase in CTexploreR, which is 52.2727273 % of our database.

75.0275028% of these genes are not testis specific.

However 40 of these lost genes are flagged as Cancer-Testis preferential in our analysis.

90.3190319 % are not properly activated in tumors and/or cancer cell lines.

plot(Carter_ours, gp = gp)

Carter_lost <- all_genes %>%
  filter(external_gene_name %in% Carter_ours@IntersectionSets[["10"]])

table(Carter_lost$testis_specificity)
## 
## not_testis_specific testis_preferential     testis_specific 
##                   1                  22                  34
table(Carter_lost$CT_gene_type)
## 
## CTP_gene    other 
##        6       51
table(Carter_lost$not_detected_in_somatic_HPA)
## 
## FALSE  TRUE 
##    13    44
table(Carter_lost$TCGA_category)
## 
##     activated         leaky not_activated 
##            28            14            15
table(Carter_lost$CCLE_category)
## 
##     activated         leaky not_activated 
##            19             3            35
table(Carter_lost$TCGA_category, Carter_lost$CCLE_category)
##                
##                 activated leaky not_activated
##   activated             9     0            19
##   leaky                 6     3             5
##   not_activated         4     0            11
Carter_lost_upset <- 
  list(`Not testis specific` = 
         filter(Carter_lost,
                testis_specificity != "testis_specific")$external_gene_name,
       `Not tumour activated` = 
         filter(Carter_lost,
                (TCGA_category != "activated" &
                  TCGA_category != "multimapping_issue")|
                   CCLE_category != "activated")$external_gene_name,
       `CT preferential` =
         filter(Carter_lost,
                CT_gene_type == "CTP_gene")$external_gene_name)

upset_Carter <- fromList(Carter_lost_upset)
upset(upset_Carter,
      text.scale = upset_text_size)

We find 38.8349515 % of CTdatabase in CTexploreR, which is 22.7272727 % of our database.

40.3508772% of these genes are not testis specific.

However 6 of these lost genes are flagged as Cancer-Testis preferential in our analysis.

84.2105263 % are not properly activated in tumors and/or cancer cell lines.

plot(Bruggeman_ours, gp = gp)

Bruggeman_lost <- all_genes %>%
  filter(external_gene_name %in% Bruggeman_ours@IntersectionSets[["10"]])

table(Bruggeman_lost$testis_specificity)
## 
## not_testis_specific testis_preferential     testis_specific 
##                 627                  61                  10
table(Bruggeman_lost$CT_gene_type)
## 
## CTP_gene    other 
##        6      692
table(Bruggeman_lost$not_detected_in_somatic_HPA)
## 
## FALSE  TRUE 
##   441   233
table(Bruggeman_lost$TCGA_category)
## 
##          activated              leaky multimapping_issue      not_activated 
##                 45                624                 11                 18
table(Bruggeman_lost$CCLE_category)
## 
##     activated         leaky not_activated 
##           242           428            28
table(Bruggeman_lost$TCGA_category, Bruggeman_lost$CCLE_category)
##                     
##                      activated leaky not_activated
##   activated                 32     3            10
##   leaky                    193   418            13
##   multimapping_issue         9     0             2
##   not_activated              8     7             3
Bruggeman_lost_upset <- 
  list(`Not testis specific` = 
         filter(Bruggeman_lost,
                testis_specificity != "testis_specific")$external_gene_name,
       `Not tumour activated` = 
         filter(Bruggeman_lost,
                (TCGA_category != "activated" &
                  TCGA_category != "multimapping_issue")|
                   CCLE_category != "activated")$external_gene_name,
       `CT preferential` =
         filter(Bruggeman_lost,
                CT_gene_type == "CTP_gene")$external_gene_name)

upset_Bruggeman <- fromList(Bruggeman_lost_upset)
upset(upset_Bruggeman,
      text.scale = upset_text_size)

We find 1.7195767 % of CTdatabase in CTexploreR, which is 7.3863636 % of our database.

98.5673352% of these genes are not testis specific.

However 6 of these lost genes are flagged as Cancer-Testis preferential in our analysis.

95.4154728 % are not properly activated in tumors and/or cancer cell lines.

2.5 Characterisation of differences with all databases

common <- unique(c(core_ours@IntersectionSets[["11"]], 
                   CTdatabase_ours@IntersectionSets[["11"]], 
                   Wang_ours@IntersectionSets[["11"]], 
                   Carter_ours@IntersectionSets[["11"]],
                   Bruggeman_ours@IntersectionSets[["11"]]))

length(common)
## [1] 119
length(common)/dim(CT_genes)[1] * 100
## [1] 67.61364
lost_list <- unique(c(core_ours@IntersectionSets[["10"]],
                      CTdatabase_ours@IntersectionSets[["10"]],
                      Wang_ours@IntersectionSets[["10"]],
                      Carter_ours@IntersectionSets[["10"]],
                      Bruggeman_ours@IntersectionSets[["10"]]))

lost <- all_genes %>%
  filter(external_gene_name %in% lost_list)

all_lost_upset <- 
  list(`Not testis specific` = 
         filter(lost,
                testis_specificity != "testis_specific")$external_gene_name,
       `Not tumour activated` = 
         filter(lost,
                (TCGA_category != "activated" &
                  TCGA_category != "multimapping_issue")|
                   CCLE_category != "activated")$external_gene_name,
       `CT preferential` =
         filter(lost,
                CT_gene_type == "CTP_gene")$external_gene_name)

upset_all <- fromList(all_lost_upset)
upset(upset_all,
      text.scale = upset_text_size)

not_specific <- filter(lost, testis_specificity == "not_testis_specific")

GTEX_expression(not_specific$external_gene_name, units = "log_TPM")
## Warning: 5 out of 988 names invalid: FIRRM, CFAP96, SPMIP5, LIAT1, BLTP2.
## See the manual page for valid types.

somatic_testis <- filter(lost, not_detected_in_somatic_HPA == FALSE)

testis_expression(somatic_testis$external_gene_name, cells = "all")
## Warning: 11 out of 710 names invalid: MPL, EDAR, ASB14, F2RL2, PCDHGB2, FOXO3B,
## KRT33B, SEMG1, ADORA2A, DAZ2, DAZ4.
## See the manual page for valid types.
## `use_raster` is automatically set to TRUE for a matrix with more than
## 2000 columns You can control `use_raster` argument by explicitly
## setting TRUE/FALSE to it.
## 
## Set `ht_opt$message = FALSE` to turn off this message.
## 'magick' package is suggested to install to give better rasterization.
## 
## Set `ht_opt$message = FALSE` to turn off this message.

HPA_cell_type_expression(somatic_testis$external_gene_name)

not_TCGA_activated <- filter(lost, TCGA_category != "activated" & 
                               TCGA_category != "multimapping_issue")

TCGA_expression(not_TCGA_activated$external_gene_name,
                tumor = "all",
                units = "log_TPM")
## Warning: 16 out of 1283 names invalid: FIRRM, CIMIP2C, CFAP96, SPMIP10, SPMIP4,
## SPMIP7, SPATA31F1, SPMIP5, CIMAP1A, SAXO4, CIMAP1C, LIAT1, BLTP2,
## SPMAP2, CIMAP1D, SAXO5.
## See the manual page for valid types.
## `use_raster` is automatically set to TRUE for a matrix with more than
## 2000 columns You can control `use_raster` argument by explicitly
## setting TRUE/FALSE to it.
## 
## Set `ht_opt$message = FALSE` to turn off this message.
## 'magick' package is suggested to install to give better rasterization.
## 
## Set `ht_opt$message = FALSE` to turn off this message.

not_CCLE_activated <- filter(lost, CCLE_category  != "activated")

CCLE_expression(not_CCLE_activated$external_gene_name,
                  type = c("lung", "skin", "colorectal",
                           "gastric", "breast", "head_and_neck"),
                units = "log_TPM")
## Warning: 11 out of 1031 names invalid: FIRRM, SPMIP10, SPMIP4, SPATA31F1,
## CIMIP2A, CIMAP1A, SAXO4, CIMAP1C, LIAT1, BLTP2, CIMAP1D.
## See the manual page for valid types.

transcript_prob <- lost %>% 
  filter(testis_specificity == "testis_specific" |
           testis_specificity == "testis_preferential") %>%
  filter(TCGA_category == "activated" | TCGA_category == "multimapping_issue") %>% 
  filter(CCLE_category == "activated") %>% 
  dim()

119 genes in our CTexploreR database are found in at least one of the other database, which represents 67.6136364%.

We have lost 1695 genes in total. Among them, 58.2890855% are not considered testis specific, 41.8879056% are expressed in somatic cells, 75.6932153% are not activated in TCGA samples, 60.8259587% are not activated in CCLE cell lines and 3.4218289% is lost due to transcripts problems.

What about new genes in CTexploreR

new <- CT_genes %>% 
  filter(!external_gene_name%in%common)

new
table(new$testis_specificity)
## 
## testis_specific 
##              57
table(new$X_linked)
## 
## FALSE  TRUE 
##    46    11
table(new$regulated_by_methylation)
## 
## FALSE  TRUE 
##    23    34
table(new$X_linked, new$regulated_by_methylation)
##        
##         FALSE TRUE
##   FALSE    23   23
##   TRUE      0   11
TCGA_expression(tumor = "all", genes = new$external_gene_name, 
                units = "log_TPM")
## `use_raster` is automatically set to TRUE for a matrix with more than
## 2000 columns You can control `use_raster` argument by explicitly
## setting TRUE/FALSE to it.
## 
## Set `ht_opt$message = FALSE` to turn off this message.
## 'magick' package is suggested to install to give better rasterization.
## 
## Set `ht_opt$message = FALSE` to turn off this message.

TCGA_expression(tumor = "all", 
                genes = filter(new, X_linked & regulated_by_methylation)$external_gene_name, 
                units = "log_TPM")
## `use_raster` is automatically set to TRUE for a matrix with more than
## 2000 columns You can control `use_raster` argument by explicitly
## setting TRUE/FALSE to it.
## 
## Set `ht_opt$message = FALSE` to turn off this message.
## 'magick' package is suggested to install to give better rasterization.
## 
## Set `ht_opt$message = FALSE` to turn off this message.

There are 57 new CT genes in CTexploreR. These are all testis specific and mainly on autosomes. Regulation by methylation is the majority of them. There is only 11 new “major” CT that are on the X chromosome and regulated by methylation. CT45 are not that new.

Expression in tumours doesn’t strike that much.

3 SessionInfo

sessionInfo()
## R version 4.4.0 (2024-04-24)
## Platform: x86_64-pc-linux-gnu
## Running under: Ubuntu 22.04.4 LTS
## 
## Matrix products: default
## BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3 
## LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.20.so;  LAPACK version 3.10.0
## 
## locale:
##  [1] LC_CTYPE=C.UTF-8       LC_NUMERIC=C           LC_TIME=C.UTF-8       
##  [4] LC_COLLATE=C.UTF-8     LC_MONETARY=C.UTF-8    LC_MESSAGES=C.UTF-8   
##  [7] LC_PAPER=C.UTF-8       LC_NAME=C              LC_ADDRESS=C          
## [10] LC_TELEPHONE=C         LC_MEASUREMENT=C.UTF-8 LC_IDENTIFICATION=C   
## 
## time zone: Europe/Brussels
## tzcode source: system (glibc)
## 
## attached base packages:
## [1] stats4    stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] SingleCellExperiment_1.26.0 UpSetR_1.4.0               
##  [3] SummarizedExperiment_1.34.0 Biobase_2.64.0             
##  [5] GenomicRanges_1.56.0        GenomeInfoDb_1.40.0        
##  [7] IRanges_2.38.0              S4Vectors_0.42.0           
##  [9] BiocGenerics_0.50.0         MatrixGenerics_1.16.0      
## [11] matrixStats_1.3.0           lubridate_1.9.3            
## [13] forcats_1.0.0               stringr_1.5.1              
## [15] dplyr_1.1.4                 purrr_1.0.2                
## [17] tidyr_1.3.1                 tibble_3.2.1               
## [19] ggplot2_3.5.1               tidyverse_2.0.0            
## [21] biomaRt_2.60.0              Vennerable_3.1.0.9000      
## [23] CTexploreR_1.0.0            CTdata_1.4.0               
## [25] readr_2.1.5                 readxl_1.4.3               
## 
## loaded via a namespace (and not attached):
##   [1] DBI_1.2.2               RBGL_1.80.0             gridExtra_2.3          
##   [4] httr2_1.0.1             rlang_1.1.3             magrittr_2.0.3         
##   [7] clue_0.3-65             GetoptLong_1.0.5        compiler_4.4.0         
##  [10] RSQLite_2.3.6           reshape2_1.4.4          png_0.1-8              
##  [13] vctrs_0.6.5             pkgconfig_2.0.3         shape_1.4.6.1          
##  [16] crayon_1.5.2            fastmap_1.1.1           dbplyr_2.5.0           
##  [19] XVector_0.44.0          labeling_0.4.3          utf8_1.2.4             
##  [22] rmarkdown_2.26          tzdb_0.4.0              graph_1.82.0           
##  [25] UCSC.utils_1.0.0        bit_4.0.5               xfun_0.43              
##  [28] zlibbioc_1.50.0         cachem_1.0.8            jsonlite_1.8.8         
##  [31] progress_1.2.3          blob_1.2.4              highr_0.10             
##  [34] DelayedArray_0.30.1     prettyunits_1.2.0       parallel_4.4.0         
##  [37] cluster_2.1.6           R6_2.5.1                stringi_1.8.4          
##  [40] bslib_0.7.0             RColorBrewer_1.1-3      jquerylib_0.1.4        
##  [43] cellranger_1.1.0        Rcpp_1.0.12             iterators_1.0.14       
##  [46] knitr_1.46              timechange_0.3.0        Matrix_1.7-0           
##  [49] tidyselect_1.2.1        rstudioapi_0.16.0       abind_1.4-5            
##  [52] yaml_2.3.8              doParallel_1.0.17       codetools_0.2-19       
##  [55] curl_5.2.1              plyr_1.8.9              lattice_0.22-6         
##  [58] withr_3.0.0             KEGGREST_1.44.0         evaluate_0.23          
##  [61] BiocFileCache_2.12.0    xml2_1.3.6              circlize_0.4.16        
##  [64] ExperimentHub_2.12.0    Biostrings_2.72.0       pillar_1.9.0           
##  [67] BiocManager_1.30.23     filelock_1.0.3          foreach_1.5.2          
##  [70] generics_0.1.3          vroom_1.6.5             BiocVersion_3.19.1     
##  [73] hms_1.1.3               munsell_0.5.1           scales_1.3.0           
##  [76] glue_1.7.0              tools_4.4.0             AnnotationHub_3.12.0   
##  [79] grid_4.4.0              AnnotationDbi_1.66.0    colorspace_2.1-0       
##  [82] GenomeInfoDbData_1.2.12 cli_3.6.2               rappdirs_0.3.3         
##  [85] fansi_1.0.6             S4Arrays_1.4.0          ComplexHeatmap_2.20.0  
##  [88] gtable_0.3.5            sass_0.4.9              digest_0.6.35          
##  [91] SparseArray_1.4.3       ggrepel_0.9.5           farver_2.1.2           
##  [94] rjson_0.2.21            memoise_2.0.1           htmltools_0.5.8.1      
##  [97] lifecycle_1.0.4         httr_1.4.7              mime_0.12              
## [100] GlobalOptions_0.1.2     bit64_4.0.5